Advanced AI Prompt for Reusable Composite Actions Framework

Act as a CI/CD Tooling Architect. Engineer a standardized GitHub Composite Action encapsulated within an `action.yml` file, named [ACTION_NAME], designed to eliminate DRY (Don't Repeat Yourself) violations across a massive enterprise organization. The action must: 1) Statically define typed [INPUT_PARAMS] (e.g., required strings, booleans with defaults) for high flexibility. 2) Encapsulate a complex, multi-step sequence of bash commands executing a proprietary security linting protocol. 3) Parse the exit codes and map the results to standardized [OUTPUT_VALS] returned to the invoking workflow. 4) Include comprehensive YAML metadata (`description`, `author`, `branding` colors/icons) formatted for publication to the internal GitHub Enterprise Marketplace. Provide an embedded code snippet demonstrating exactly how an external repository's `main.yml` would invoke this composite action.